U.S. patent number 8,045,771 [Application Number 11/986,494] was granted by the patent office on 2011-10-25 for system and method for automated patient anatomy localization.
This patent grant is currently assigned to General Electric Company. Invention is credited to Mukta Chandrashekhar Joshi, James Vradenburg Miller, Robert Franklin Senzig, Xiaodong Tao.
United States Patent |
8,045,771 |
Tao , et al. |
October 25, 2011 |
**Please see images for:
( Certificate of Correction ) ** |
System and method for automated patient anatomy localization
Abstract
A method of imaging comprises performing a scout image of an
object and aligning major regions of interest within the object by
comparing the scout image with a pre-determined atlas image. In a
further embodiment, a method of imaging using a computed tomography
(CT) scanner is provided and comprises performing a scout image of
a subject and aligning major anatomical regions of interest within
the subject by comparing the scout image with a pre-determined
atlas image. Further, an imaging system for us with a computed
tomography (CT) imaging device is provided and comprises a
processor configured to perform aligning major anatomical regions
of interest within a subject by comparing a scout image with a
pre-determined atlas image, and wherein the major anatomical
regions are used for automated localization for Scan Range and/or
Exam split.
Inventors: |
Tao; Xiaodong (Niskayuna,
NY), Miller; James Vradenburg (Clifton Park, NY), Joshi;
Mukta Chandrashekhar (Belmont, MA), Senzig; Robert
Franklin (Germantown, WI) |
Assignee: |
General Electric Company
(Niskayuna, NY)
|
Family
ID: |
39584091 |
Appl.
No.: |
11/986,494 |
Filed: |
November 21, 2007 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20080159611 A1 |
Jul 3, 2008 |
|
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
60860639 |
Nov 22, 2006 |
|
|
|
|
Current U.S.
Class: |
382/128; 382/131;
382/294; 382/293; 378/8; 382/132 |
Current CPC
Class: |
A61B
6/08 (20130101); A61B 6/488 (20130101); A61B
6/032 (20130101) |
Current International
Class: |
G06K
9/00 (20060101) |
Field of
Search: |
;382/128,131,132,154,294 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Lu; Tom Y
Attorney, Agent or Firm: Klindtworth; Jason K.
Parent Case Text
CROSS REFERENCE TO RELATED APPLICATIONS
This application is related to Provisional Application U.S. Ser.
No. 60/860,639, entitled "SYSTEM AND METHOD FOR AUTOMATED PATIENT
ANATOMY LOCALIZATION USING SCOUTS", filed Nov. 22, 2006, the
contents of which are herein incorporated by reference and the
benefit of priority to which is claimed under 35 U.S.C. 119(e).
Claims
What is claimed:
1. A method of imaging comprising: performing a scout image of an
object; aligning major regions of interest within the object by
comparing the scout image with a pre-determined atlas image to
define a limited scan range for image acquisition; acquiring image
data based on the limited scan range.
2. The method of claim 1 wherein the aligning step further
comprises calculating a similarity measurement between the scout
image and altas image.
3. The method of claim 1 further comprising identifying more
detailed or finer regions of interest contained within the object
after aligning the major regions of interest to localize detailed
regions of interest.
4. The method of claim 1 wherein the object is a patient and the
major regions of interest comprise anatomical regions of interest
comprising chest, abdomen, pelvis, and combinations thereof.
5. The method of claim 3 wherein the detailed or finer regions of
interest comprise organs and anatomical structures comprising eyes,
breasts, liver, and combinations thereof.
6. The method of claim 1 wherein the scout and atlas images are
acquired with a same computed tomography (CT) scanner geometry.
7. The method of claim 6 wherein the scout and atlas images are
acquired with one of multiple two-dimensional (2D) scouts or a low
dose helical acquisition.
8. A method of imaging using a computed tomography (CT) scanner
comprising: performing a scout image of a subject; aligning major
anatomical regions of interest within the subject by comparing the
scout image with a pre-determined atlas image to define a limited
scan range for image acquisition; acquiring image data based on the
limited scan range.
9. The method of claim 8 further comprising identifying more
detailed or finer regions of interest contained within the subject
after aligning the major anatomical regions of interest to localize
detailed regions of interest.
10. The method of claim 9 wherein the major anatomical regions of
interest comprise chest, abdomen, pelvis, and combinations
thereof.
11. The method of claim 9 wherein the detailed or finer regions of
interest comprise organs and anatomical structures comprising eyes,
breasts, liver, and combinations thereof.
12. The method of claim 8 wherein the major anatomical regions are
used for automated localization for Exam split.
13. The method of claim 8 wherein the anatomical regions are used
for automatic identification of patient orientation within a gantry
of the CT scanner.
14. The method of claim 8 further comprising: using information
corresponding to the major anatomical regions of interest for
automated post processing steps comprising anatomy based reformat,
anatomy based context sensitive processing to extract relevant
information, Computer Aided detection of pathology and combinations
thereof.
Description
BACKGROUND
In the current computed tomography (CT) image acquisition process,
X-Ray scout images are acquired for anatomy localization. An
operator reviews the scouts and sets scanning parameters to acquire
images of patient region of interest. Patient anatomy is unknown to
the scanner. Furthermore, after the images have been reconstructed,
typically the operator has to divide the scanned images into
anatomical regions to bill and network to the appropriate
department within radiology e.g. Chest, Abdomen etc. This is called
Exam Split.
The annotation on the images depends on the patient's orientation
within the gantry at the time of scanning. This orientation
(whether head first or feet first etc) is manually entered by the
operator. There have been times when this orientation is either
entered incorrectly or due to other factors, the patient
orientation was changed after the orientation on the scanner
prescription was entered. This causes the annotation of the patient
images to be incorrect which can lead to serious consequences.
There is a need for an automated process of Exam Split. Further
there is a need for the scanner to be able to identify anatomy, and
therefore it can also identify the orientation of the patient when
scanning and the annotation of the images will be directly
correlated to the actual scanned data, thus eliminating errors as
described above.
BRIEF DESCRIPTION
In a first aspect, a method of imaging comprises performing a scout
image of an object and aligning major regions of interest within
the object by comparing the scout image with a pre-determined atlas
image.
In a second aspect, a method of imaging using a computed tomography
(CT) scanner is provided and comprises performing a scout image of
a subject and aligning major anatomical regions of interest within
the subject by comparing the scout image with a pre-determined
atlas image.
In a third aspect, an imaging system for us with a computed
tomography (CT) imaging device is provided and comprises a
processor configured to perform aligning major anatomical regions
of interest within a subject by comparing a scout image with a
pre-determined atlas image, and wherein the major anatomical
regions are used for automated localization for Scan Range and/or
Exam split.
DRAWINGS
These and other features, aspects, and advantages of the present
invention will become better understood when the following detailed
description is read with reference to the accompanying drawings in
which like characters represent like parts throughout the drawings,
wherein:
FIG. 1 is a schematic illustration of a registration framework to
which embodiments of the present invention are applicable;
FIG. 2 is an exemplary illustration of images obtained using
methods of the present invention; and,
FIG. 3 is an exemplary illustration of images obtained using
methods of the present invention.
DETAILED DESCRIPTION
This invention solves the problem of automatic localization of
patient anatomy. This can be done from one or more scout images or
from a 3D dataset acquired from a low dose helical "scout". For
example, in a typical CT imaging session, two scout images are
acquired first. An operator reviews the scouts and sets scanning
parameters to acquire images of patient region of interest. This
invention automates this process by automatically finding the
patient anatomy from the scout image(s) (2D or 3D). This also
allows for automation of the Exam Split process thus improving the
workflow and increasing the efficiency.
In embodiments of the present invention, there is provided a system
and method to automatically localize patient anatomy from scout
images (one or more 2D scouts or a low dose helical scout) by
registering a pre-labeled atlas into the patient space and
transforming the atlas labels to find patient anatomy. The system
can also recognize the patient orientation. The system can be used
in medical imaging devices for automatic image range prescription,
more accurate patient-centric dose management, automatic exam
split, as well as for correct identification of patient orientation
and for automated reconstruction parameter optimization per
anatomy.
Modern medical imaging devices can acquire patient images with
better quality in shorter time. What they see, however, is nothing
more than a collection of numbers. The accurate patient anatomy is
completely unknown to the scanner. Therefore, an operator's
involvement is often required to figure out the patient position
and orientation from a number of scout images and to set scanning
parameters appropriate for the study to be conducted. Moreover,
since the patient anatomy is unknown to the medical imaging
devices, dose management can only be implemented in a device
centric way.
In order to tackle the problem of anatomy localization from scouts,
a framework for registering a 3D volumetric atlas to one or more
scout images is provided. The resulting transformation is then used
to transform the atlas segmentation (anatomical labels) to the
patient space to localize patient anatomy.
Referring to FIG. 1, architecture for this registration framework
is shown. In the step of aligning atlas to the scout images,
simulated scout images of a spatially transformed version of the
atlas are generated with the same scanner geometry. The simulated
images are compared with the patient scout images for a similarity
measurement, which is then used to update the transform to improve
the alignment between the atlas and the patient.
To get an accurate estimation of the position/orientation of
patient organ/region of interest, a global to local search strategy
is employed that aligns major body sections first, and switches to
local regions to localize detailed patient anatomy. In one
embodiment one could use the Single metric as Mutual Information,
the Many-to-One metric as a weighted sum of the single metrics, the
optimizer could be one does not require gradient information e.g.
Amoebae style optimizer. Those skilled in the art would recognize
that a variety of metric could be used to measure the similarity as
well as a variety of optimization techniques could be used to
estimate the parameters of the transformation. The Scout generator
uses the scanner geometry and the current estimation of the
transformation to generate the hypothesized Atlas Scout Image. See
example result in FIG. 2. Referring now to the FIG. 3 shows the
transfer of the labels from the atlas to the target. Initially
without registration the label was in the wrong location, after the
registration, the label is seen to be in the correct location on
the target.
In a first embodiment, a method of imaging comprises performing a
scout image of an object and aligning major regions of interest
within the object by comparing the scout image with a
pre-determined atlas image. The aligning step further comprises
calculating a similarity measurement between the scout image and
atlas image. Further, the method comprises identifying more
detailed or finer regions of interest contained within the object
after aligning the major regions of interest to localize detailed
regions of interest.
In embodiments of the present invention, the object is a patient
and the major regions of interest comprise anatomical regions of
interest comprising chest, abdomen, pelvis, and combinations
thereof. The detailed or finer regions of interest comprise organs
and anatomical structures comprising eyes, breasts, liver, and
combinations thereof. The scout and atlas images are acquired with
a same computed tomography (CT) scanner geometry. In a further
embodiment, the scout and atlas images are acquired with one of
multiple two-dimensional (2D) scouts or a low dose helical
acquisition.
In another embodiment, a method of imaging using a computed
tomography (CT) scanner comprises performing a scout image of a
subject and aligning major anatomical regions of interest within
the subject by comparing the scout image with a pre-determined
atlas image. The anatomical regions are used for automated
localization for Scan Range and/or Exam split. Further, the
anatomical regions are used for automatic identification of patient
orientation within a gantry of the CT scanner. In a further
embodiment, the method comprises using information corresponding to
the major anatomical regions of interest for automated post
processing steps comprising anatomy based reformat, anatomy based
context sensitive processing to extract relevant information,
Computer Aided detection of pathology and combinations thereof.
In a further embodiment, an imaging system for us with a computed
tomography (CT) imaging device is provided and comprises a
processor configured to perform aligning major anatomical regions
of interest within a subject by comparing a scout image with a
pre-determined atlas image, and wherein the major anatomical
regions are used for automated localization for Scan Range and/or
Exam split. Further, the anatomical regions are used for automatic
identification of patient orientation within a gantry of the CT
imaging device. In a further embodiment, the processor is further
configured for using information corresponding to the major
anatomical regions of interest for automated post processing steps
comprising anatomy based reformat, anatomy based context sensitive
processing to extract relevant information, Computer Aided
detection of pathology and combinations thereof.
It is to be appreciated that there are a number of applications and
uses for the present invention, for example: The method of
identifying gross body parts or organ locations by mapping
information to pre-scan imagery in the form of multiple 2D scouts
or a low dose helical acquisition. The method of identifying gross
body parts or organ locations by mapping information from an
anatomical atlas to pre-scan imagery in the form of multiple 2D
scouts or a low dose helical acquisition. Automatic identification
of gross body areas e.g. Chest, abdomen, pelvis etc. Automatic
identification of more detailed organs and structures e.g. eyes,
breasts, liver etc Use of anatomy information for dose management
to individual structures Use of anatomy information for automated
localization for Scan Range, Exam split Use of anatomy information
for automated post processing steps including but not limited to
anatomy based reformat, anatomy based context sensitive processing
to extract relevant information, Computer Aided detection of
pathology Automatic optimization of reconstruction parameters per
anatomy e.g. recon kernels, display field of view which affects
resolution, image thickness etc. Use of anatomy information for
automatic identification of patient orientation within the Scanner
gantry. This can ensure correct annotation of the images without
relying on operator input and reducing errors caused by incorrect
input. Use of Anatomy information to automatically assign correct
preset for Window & Level, which can be used in a pure CT as
well as a PET/CT system.
This has broad applications through out the system from acquisition
protocol to post processing and in overall workflow through
automation of different steps.
While only certain features of the invention have been illustrated
and described herein, many modifications and changes will occur to
those skilled in the art. It is, therefore, to be understood that
the appended claims are intended to cover all such modifications
and changes as fall within the true spirit of the invention.
* * * * *